Dr. Tamar Mitts (2:56)

Thank you so much for having me.

Dr. Miranda Melcher (2:57)

Could you please start us off by introducing yourself a little bit and tell us why you decided to write this book, what questions you're asking in it, how this all developed to get to this point?

Dr. Tamar Mitts (3:07)

Yes. Thank you. Yeah. So I'm an associate professor at Columbia School of International and Public Affairs. I have been really interested in the information environment writ large. And over the past, I would say, probably 15 years, I've been studying, you know, different aspects of the online environment, particularly social media platforms. And one of the things I became really fascinated with is how difficult it has been to address issues that I think many people around the world would agree are problematic. So I'm talking about extreme hate or promotion of violence or discrimination across different lines. And I was really curious why we as societies have such hard time to deal with this. And it seems to be, on one hand, an old issue and an old problem, but on the other hand, it seems to be kind of raising new questions on, like, what to do in a world where we are more connected than ever before. So ideas flow much more faster, which I think is a great thing on one hand. But on the other hand, it actually creates new problems. And to me, those new problems are really, really puzzling and interesting. So that was kind of the background of why I chose to tackle this particular question, which admittedly is a really difficult challenge.

Dr. Miranda Melcher (4:35)

I mean, tricky. Definitely a big challenge, but someone has to investigate it. So I'm glad you've taken it on. Beyond the sort of big picture questions of sort of, why is this such a big problem and to what extent is it a new problem? Are there any other questions you're asking in the book that you want to lay out to begin our discussion?

Dr. Tamar Mitts (4:54)

Yes, sure. So kind of nailing this a little bit more narrow, what I focused on in the book is specifically the issue of extremism. And I think we'll probably have a chance to discuss a bit in length. What is extremism? Kind of how we should think about extremism, but kind of, to put it in a short way, I think of extremist groups are those that are defined by state actors by the international community as those groups that promote mostly violent ideologies and as a result are often banned from social media platforms. And to me, that was really interesting to study these groups. And I've been studying them for many years, looking at how they use the online space for recruitment, to radicalize potential supporters, to raise funds. And what was so interesting to me as I was studying them is that I saw two things happening at the same time. For one, I saw that many tech companies facing pressure, oftentimes by governments, actually kind of address that issue on their platform and trying to ban these accounts, ban their networks and take them off their services. But on the other hand, I saw these groups become really creative in how they overcome those bans and what they do in response. So I was really motivated to document that particular adaptation that these groups are doing. And then while I was studying this, I saw that there was like, omg, there is like a whole ecosystem here. And it's a really multi platform phenomena. And the way we think about the problem is not in a multi platform perspective. So my motivation for writing this book, kind of focusing more specifically on extremism or violent extremism, was that we are kind of looking at the problem in the wrong way. And that's kind of the underlying motivation for the book is to explain why and perhaps how we should be thinking about it. Even though I admit I don't have the perfect solution either. But I think studying this and understanding the problem maybe could help us solve the problem eventually.

Dr. Miranda Melcher (7:06)

Yeah, we definitely can't solve something if we don't know what we're talking about. So we should probably make sure that we define some terms then early in our conversation to make sure we don't end up confusing ourselves later on. So extremism and harmful content, what precisely do we mean by those terms in this context?

Dr. Tamar Mitts (7:25)

Yes, so this is something I started, I kind of start the book with mainly because as I'm sure you and our listeners know, these have become very politicized and controversial terms. So it's kind of like today one can say that if you don't like somebody's ideas, you can call them extremists, or if you don't like something or disagree with someone, you can say that their ideas are harmful. And that is not really what I mean in the book. So in the book I draw on different traditions of scholarship that look at this question of extremism and I come up to the following definition that draws on these things. So on one hand, extremism could be seen as Kind of at the idea level. So ideas that deviate from the majority of the population. However you look at a population, it could be a country, you know, a region, a global community. And these would be kind of defined in relation to the majority, which could be something bad or not bad. But it's just kind of looking at ideas and how they are distributed across a population. Something on the extreme ends of a spectrum you can define perhaps as extremists. So this is kind of one way to think about extreme. Another one, which is a bit more common, is to look at extremism in the context of intergroup relations. So in any society, we have different societal groups that may be in tension with each other. So in that context, extremism would be ideas or actions that promote, in one way or another, harm to the other group that you're in tension with. And that really links to the third way to think about extremism, which relates to violence. And a lot of people in the space of extremism oftentimes conflate extremism and violent extremism as this sort of the same thing. And that's actually kind of on purpose is, I think, driven by this idea of trying to tackle a problem that many would agree is a problem. So, like, hurting people physically in the real world is not something many societies would want. So therefore, extremist groups or extremist actors would be those that promote violence more specifically. So in the book, I look at all of these different traditions, and I say, okay, I don't want to say here myself what is good and what is bad. This is not my business. But what I really am interested is understanding how different societies define extremism. Who are the actors that are then defined as extremists? Those are. Those are the groups that I study. But I would say that there is also a twist here, because even though you can think of, say, okay, I think every actor that promotes violence against civilians, say, is an extremist actor. Well, that also varies over time. And the definition of what is extremist also has evolved over time. And even in the course of writing the book, the same actors have changed their definition of what is extremist throughout the time I was studying this phenomena. So I think it's a very fluid concept as a result. And that's why each time I talk about extremism or talk about this book, I always try to caveat and say, I'm not coming here with a normative stand. But I just want to examine, once we tag an actor as extremist how can we understand their behavior in that context? So I know it might be a little bit of a long answer to your question, but I think that it's important to clarify that the goal of this book is not to come with a normative statement about who is extremist and who is not, but to really study the concept of extremism and how it's applied in the context of content moderation. Yeah.

Dr. Miranda Melcher (11:13)

Definitions are so key. So I'm really glad that you explained that fully so that we can have a firm basis to have the rest of the discussion on now that we know what we're talking about. So I think then it's worth starting to get into some of the things that have been done once these tags have been put on of extremists, what that has led to in the past, to kind of assess how that's gone and what might happen in the future. So can you give us a sense of the different ways that democratic governments have regulated extremist behavior on social media? What are sort of the range of options or tools that they've tried?

Dr. Tamar Mitts (11:51)

Yes, and that's something that I learned a lot in the course of the book, because I thought, I'm here in the US And I am very familiar with the United States approach to these questions. And as I was studying other countries, I was amazed to see how many countries actually have studied, tried to deal with the same problem, but in different ways. So I should say that, yes, you mentioned democratic governments. I definitely focus on democratic governments in the book, particularly because once you move into less democratic settings, the assumptions around speech, for example, start to differ a lot. For example, if you think of authoritarian settings and kind of who is defined as extremist, that kind of starts changing a lot of the assumptions that I kind of operate with in the book. So I should say that when I talk about speech regulation, I talk about this from the perspective of countries that at baseline, would want to have liberal values of freedom of speech for everybody. And then when they are faced with a challenge of actors that are promoting extremism, again, going off of our definition of what they see is unacceptable in their societies, then they have this challenge of, okay, where do we draw the line between how much speech is allowed and how much should be regulated? And there's a whole legal literature around these questions specifically. But what I was looking for in the book specifically is to say, okay, so different countries, different democratic countries are dealing with this challenge of growing hate and extremism on their online communities. What do they do about it and what I found is that many tend to focus on the platforms that they think are most important. So that tends to be the platforms that everybody knows that are the largest and most familiar, precisely because that's perhaps where most people would see extreme rhetoric or extreme promotion of violence against other people as opposed to other online spaces. So what you see country after country, even with varying assumptions around how much speech should be regulated, because there is variation, what you see in all the cases, pretty much, I would say almost all the cases. And I looked at all the countries, many, many different countries in my data, they basically tend to differentiate platforms based on how big they are. So usually there's going to be some threshold of, of like monthly active users, for example, or some other threshold of platform size above which a platform would face many regulations if it doesn't abide by the country's rules around hateful content or extremist speech. And below it, many platforms would kind of face either more limited regulation or no regulation at all. And that's actually quite interesting because it's. I argue in the book, it actually is an important background factor for understanding why then different platforms have different standard standards. Sorry. For moderating such content. And that actually creates the phenomena that I later study in the book. So just to say that we have this really interesting dilemma that countries are dealing with, which many have come to think of as, okay, we don't want to limit all speech, but we're going to focus on the actors that we think matter the most, and that actually has maybe some unintended consequences.

Dr. Miranda Melcher (15:27)

Yeah, that's definitely helpful to point out. And point out as not just being, oh, this is an interesting fact. It's like, well, hang on a second, what does this actually mean? And it's not just a question of different governments acting in different ways. As you've mentioned, different social media companies also are making decisions here, and that turns out to be really crucial in the data that you collect and analyze. So can you help us understand, again, kind of what the impact of that is? So if we've got different social media companies and they each have their own policies and approaches to moderation, even if they're all trying to tackle the same problem, they're doing it in different ways. How does the fact that they're doing it in different ways impact what extremist groups can and cannot do?

Dr. Tamar Mitts (16:12)

Yes, so that's exactly the heart of the book. And I should say that the expectation is not that all platforms should tackle this exactly in the same way, precisely because we do want to give these actors the freedom to choose how they want to moderate. And that's really kind of rooted in the United States section 230. And we can discuss this at length if you want. But essentially the idea is not that we have to have a uniform standard, but to kind of argue that maybe we should think of a more uniform standard. So first I would say that secondly, different platforms have different user bases. They have different kind of affordances and tools that they allow for their users to use. And maybe some set of rules or regulations for one platform would not really make sense for another one. So there is different reasons for why we see those variations, and some variation obviously is expected. One of the things I point to, however, is variation that is kind of stark on issues that we would expect to see more kind of unified approach, if you will. And again, that's speaking directly to how we should think of moderating calls for violence or glorifying violence, or glorifying those sort of language. Now, when you have those different standards that different platforms have, whether they come from the pressure that companies face from governments or whether they come from another source, the consequence is that you have this sort of uneven moderation ecosystem. And it's not really as arbitrary as you may think. Perhaps because of all these pressures, it actually tends to look pretty predictable in the way that moderation looks like across platforms. So if you actually measure this, you would see that pretty consistently the larger platforms are more restrictive in their policies to regulate harmful content. So they would have more rules, they would have more definitions, that more would kind of fall under the category of extremism or harmful content for those platforms. And then as you go down the kind of monthly active user base or the size of the platform in whatever way you want to measure it, then the less rules you see, and that sort of relationship creates a super fascinating environment to study how extremist actors adapt. And what I study in the book is really to show how they navigate this uneven ecosystem in a way that allows them to overcome those restrictions. And I kind of go through different mechanisms by which they do this in the book, but really it's all rooted in this sort of uneven nature where large platforms have more restrictive rules and smaller platforms have more lenient rules.

Dr. Miranda Melcher (19:14)

I definitely want to get into some examples of what that looks like. So one of the things you talk about in the book is to stay on the big mainstream social media platforms. One strategy that some groups use is to change their messaging so that they don't can stay there and don't get kicked Off. Can you walk us through what that looks like? Like what are some examples of that? Does it actually work? What are sort of the impacts of the group sort of trying to get around the rules in this way?

Dr. Tamar Mitts (19:42)

Yes, that's a great question. And basically, if you think of kind of shifting now to the point of view of these groups, what do they want to do online? And in the book I describe this a little bit, kind of in a simplified way, but I think the core idea is kind of travels to many groups and I showed this in the case studies. But basically, if you think, you know, you're a group that is tagged as extremists, therefore you face these restrictions, what is your goal online? Many groups goal is to reach audiences, to attract supporters, to attract resources, et cetera. And where would they find most of these? Or like, I guess, on average, where do we find more of those is in the platforms that have more people on them. Right. So those audiences larger. That's where you want to be. But if those platforms are also more restrictive, then they need to find ways to reach those people without getting caught. And as a result, one of the strategies that I document in the book that is quite creative, I would say, is to find different ways to stay kind of below the threshold that would flag them for kind of banning or suspension. So it can look like different in different ways for different groups. But just to give some examples, some of the actors that I study in the book, they figured out that the big companies use classifiers to identify particular keywords that they're using that then kind of flag them for suspension.

Wayfair Advertiser (21:14)

The holidays are coming up and I have relatives and friends visiting, and that means I need to put them up happily. I have a couple of guest rooms. Unfortunately, these guest rooms need help and I go for help to Wayfair because Wayfair has everything that you need to spruce up your guest rooms and also your bathrooms and so on and so forth. Wayfair is really your one stop shop for all of these things. Bedding and linens and throw pillows, accent chairs, towels, cookware, plates, utensils. You name it, they have it. And they can get it to you quickly and free, because shipping is free at Wayfair. Now, I personally bought a couple of comforters and some pillows for my guest rooms, and it was extraordinarily easy to order them. They came in a few days and as I say, shipping was free. So this saved me time and I just had a great experience. So Wayfair makes it easy to tackle your home goals. This holiday with endless inspiration for every space and budget, free and easy delivery, all your seasonal must haves. So get organized, refreshed and ready for the holidays. For way less head to Wayfair.com right now to shop all things home. That's W A Y-F A I R.com Wayfair Every style, every home.

Windows Copilot Advertiser (22:23)

Meet the computer you can talk to with Copilot on Windows Working, creating and collaborating is as easy as talking. Got writer's block? Share your screen with Copilot Vision to help spark inspiration and use Copilot voice to have a conversation and brainstorm ideas. Or maybe you need some tech help with Copilot Vision or Copilot sees what you see. Let Copilot talk you through step by step guidance so you can master new apps, games and skills faster. Try now@windows.com copilot and what they started.

Dr. Tamar Mitts (22:52)

Doing is to play around with the actual characters that they're using to say what they want to say, to kind of evade that, right? Others have used if they want to share an image or a video that has a logo that again is kind of tagged for suspension, they might disguise the logo under something that looks very not problematic, for example, like a Netflix logo or an Amazon logo or something like that, or even add sort of noise to the images so that the pixels don't match to databases that companies use to identify their content. So I think what you see here is pretty quick learning of how platforms do their content moderations, those that do it, those that are very active in it, and then finding creative ways to evade it. Now I want to say why do they even bother, right? Like, why go through this effort if there's other platforms they could use to do what they want to do? I show that what they really want is really to be on these platforms. So it's kind of worth going through all of these efforts to try to evade the classifiers so that they can stay on these platforms. So it's kind of an interesting trade off on their end as well that you see that I document in the book using different cases.

Dr. Miranda Melcher (24:13)

Yeah, that's definitely interesting that clearly there are some groups where that is worth it. There are also, however, some groups that decide not to try and do that and instead to leave these big platforms that have more rules and go somewhere with fewer rules. But of course there are plenty of options in that space space too. So if a group decides that this is the path they want to go down, how do they decide which platform to migrate to? Surely it's not random, right?

Dr. Tamar Mitts (24:40)

Yeah, so this is definitely not random. And you're right, there's many different options, and I would say even growing number of options over time. We have more and more platforms, more and more spaces that you can, you know, use for this sort of activity. And in the book I kind of go back to this logic that what these groups really want is to maximize both the audience reach. So how many people would see their content and promotion materials? And the other dimension I mentioned kind of indirectly is what I call in the book authenticity, or essentially kind of how much can they post freely without kind of toning down their message like they need to do on the really restrictive platforms? And as a result, they kind of look for spaces that will allow them both to maximize as many eyeballs that can see their content on one hand, and on the other hand post as freely as they could on the other. And that kind of leads to very kind of specific part. I have this chart that I always think about when I describe this, but it's also in the book and I think chapter four, I kind of showed this empirically. But if you were to plot the platforms on kind of like X and Y, where the X axis would be platform size and the Y axis would be how lenient it is, higher values is more lenient than those kind of in the middle. So you can basically see this negative relationship between size and leniency. As I mentioned earlier, if you look at where these groups trying to go if they are banned from the big platforms, where they go is to these platforms that allow them to maximize both audience reach as well as what I call authenticity. And those tend to be very specific platforms in this sort of middle of the chart. And most familiar of those is Telegram. So Telegram has become a big hub for a lot of groups, partly, I argue, because of these more serious restrictions that groups have been starting to face on the larger platforms. So Telegram is not a small platform. It's not as large as the big ones, but it still allows quite significant audience reach. And that audience reach in fact has improved since September 2015. It's already 10 years ago. It's crazy to say, but at that time Telegram has shifted from being just kind of like almost like a two person chat sort of platform to a platform that allows broadcasting messages to many, from one to one to one to many. And that has become really attractive for a lot of these groups. And what we see is many migrating to platforms like Telegram after facing bans on those larger platforms.

Dr. Miranda Melcher (27:39)

I mean, that definitely makes sense from their point of view, not arbitrary at all. So thank you for helping us understand the law logic. When you talk about those case studies and data that you've put together to illustrate, as you said, the kind of XY of the decision making, do we see any differences depending on the ideologies of the groups? For example, groups that their extremist beliefs are over in this area versus that area make different decisions about whether to stick with the big ones or whether to go to Telegram or somewhere else or does. Is that not actually part of the decision making calculus?

Dr. Tamar Mitts (28:15)

Yeah. So you definitely see some variation. The groups that are most available to study that I have in the, in the book mostly are those that are either on the far right end of the political spectrum, so white nationalist groups, other far right organizations on one end, and on the other hand, I have groups that promote jihadi ideologies mostly. And these groups, these types of ideologies are extremely different from each other in many, many ways. And the platforms that they seek are also somewhat different. So for example, there are certain online spaces that are kind of most popular among far right groups. And you won't really see a lot of jihadi groups on them and vice versa. However, even though the ideologies of these groups are different and many groups try to recruit from different populations and have different objectives, what's surprising is that the way they choose spaces, online spaces to operate in is strikingly similar. So you would find like, you know, a far right group band. These groups actually became more of a target in like 2018, 19, 2020. And many jihadi groups were targets for content moderation a few years before. Then I would say 20, 14, 15, kind of, they. They began experiencing bands. But in both, both kinds of groups, you see them kind of going through the same process where they migrate to a platform that they think would be interesting to them. I think one example is the Proud boys trying to go to Rumble. Sorry, to, to Parlor, I think it was. And then going there and saying like, oh my gosh, there is no one here. I'm like shouting to the void and then trying to find places that, okay, where can we find audiences? Where can we find people to talk to? And then you see them kind of converging on those same type of platforms that would give them at least some audience reach, even if it's not the largest that the large platform would give them, but still are more lenient in the way they moderate their content.

Dr. Miranda Melcher (30:31)

Hmm, that's really interesting to see. Kind of where there is difference and also where there isn't. So I'm Very glad that you looked at that angle because it's definitely quite an interesting finding. Thinking then about a group that we haven't mentioned yet, we've talked a bit about governments, a bit about social media companies, a bit about different extremist groups. I wonder if we can almost narrow the focus, I suppose, to the individual level rather than the group level, because that's also something you cover in the book, is the impact of experiencing these content moderation policies, for example, on the big social media platforms. How does that relate to this story about sort of what kinds of messages extremist groups are putting out there?

Dr. Tamar Mitts (31:12)

Yes, that's. I really love how you, you look. You kind of described it from the different units that we could look and study this. So yes, one, one important aspect of it is really the individual. Right. In the end, people use platforms for different reasons and they experience their, the platform in different ways. And a really interesting experience that some users have started becoming exposed to more is having their, either their posts being deleted or their accounts being suspended or kind of, you know, at least for some period of time as a result of growing content moderation rules. And we really see, we started seeing this uptick around kind of coming up to 2020 and after that where many companies kind of upped their content moderation. We saw a lot of this happen in the US in the run up to the 2020 election. There was a lot of effort to prevent what happened in 2016 in the elections. So as a result, many big companies started becoming a lot more aggressive on kind of the content that they allow on their spaces. So as a result of this increased enforcement and more rules, individual users started feeling or seeing content moderation happening to them as well. And sometimes they would know, okay, I know why I got like banned or I know why I got this post deleted. But many times they were really surprised and they felt kind of unhappy or, you know, like something is not fair kind of. Why did they do so? I was really interested, particularly in thinking about radicalization over the individual. So what makes somebody who move from thinking, you know, one set of ideas about the world and then shifting over time to thinking about the world in a slightly different way that may promote, you know, may be considered extremist by again, these definitions we mentioned earlier and what I did for this, I was like, okay, this is really interesting. I'm going to now study a particular case. So that was the time where I saw an uptick in content moderation of the far right groups. And I knew at that time that many far Right. Groups and actors are starting to use increasingly what at the time was called alternative platforms, or essentially platforms that mimic the large platforms in the way they look and feel, but are more lenient in their rules, or as these platforms often market themselves as more free speech. So what I did was to study user activity both on the large platforms as well as on those alternative platforms. And I was even able to find the same user on kind of both platforms. And that that created a really neat empirical opportunity to look at what happens if somebody is banned on the big platform, what happens to them on these alternative platform. You can kind of study this in parallel. And I found fascinating results. So for one, I would say that many of those who got banned from the mainstream platforms became really, really angry. Like this is a really angry episode in their life. And they used the alternative platform to express that. So you could actually see the frustration from the moderation that happened to them. But then another thing that I found was really interesting and definitely relevant for the book was that at the same time I saw them starting to kind of write in threads and communities where they found people like them experiencing the same sort of bans for the same sort of beliefs that they believed in. And that's sort of what people might call echo chamber. But really this online community has started growing and at the same time different. Again, going back to the groups, different groups starting to look for supporters, found those communities on these alternative platforms to become really, really interesting to them. So you see these growing recruitment efforts by these groups in these more smaller platform alternative communities that again, I can't say for sure if people were recruited from there because that's not the data I had on the recruitment side. But I definitely saw more engagement with the content that these extremist groups then used to promote on the alternative platforms. So that all just to say that when we think about content moderation, we can maybe think about like, you know, banning an armed group, banning an extremist actor for whatever they're saying, or even an influencer or something like that. But actually there's many important impacts that are happening to the random user that may be kind of shaping also this phenomena as well. So I really wanted to give attention to the individual users in the book as well, because I think it's definitely part of the story. And also if we think about it in a cross platform way, yeah, it's.

Dr. Miranda Melcher (36:25)

Such an interesting part of this and also a really important one to remember that when we're talking about things like big picture social media company policy or government Regulation that there is the actual individual human element of this that can get obscured, but is of course so incredibly central. So having that then fresh top of mind, from your very helpful answer there, let us turn to then the kind of big picture policy side of things. Would it help, for instance, if social media companies had a more unified or consistent approach to how they moderated content?

Dr. Tamar Mitts (37:03)

Yeah, so that's a natural kind of conclusion you, you might want to come to after reading my book. And in fact, I also kind of from the again, policy conceptual level thought exactly that. I was like, oh, okay, so if all of these issues are coming out of this inconsistency we have in the online space, then maybe what we need is just more unification. And in fact, after the 2019 attack in Christchurch in New Zealand, there was a big multilateral push by governments and also international organizations joined it and tech companies joined what was called the Christchurch Call. And that was exactly that sort of effort. It was an effort where different companies would and countries would pledge to make efforts to kind of remove content that could, you know, create or inspire attacks such as the one happening in Christchurch. And if you look at the kind of details of this, there was like, obviously many things that were said and many interesting speeches said, but it's not really clear kind of how this ended up changing things practically. But the goal was, yes, definitely to create different independent organizations like the Global Internet Terrorism to counter, sorry, Global Internet Forum to Counterterrorism or the Gift CT was kind of one organization that was flagged as important there. There is another one called Tech Against Terrorism. So there was this idea to use these sort of independent actors or who were independent, I guess at that point in time already, to create a place where companies could collaborate in sharing standards, in sharing tools about how to moderate harmful content. And that was kind of the idea is to create more standardization. In fact, in reality, I don't think these efforts have succeeded as much as maybe intended and the policy kind of wins, have kind of focused elsewhere since then. But I do think that if we think of it from this sort of policy solution point of view, we definitely need to identify the problem first. And what I suggest in the book is that the way we have been thinking about the issue of extremism and violence and hate online is kind of platform by platform. So we see, you know, bad things on Facebook. Then we call out Facebook and say, you should do more. We see bad things on Twitter, which is now X. We call out, you know, you need to do more. But by doing this platform by Platform kind of policy solution approach. We're actually missing out on the problem that there needs to be kind of a more standardized solution to this, given the nature of the problem. So just a very long answer to your question. I think, yes, we do need a way to think about this in a more organized way. But at the same time, and that's the huge caveat I really want to highlight, is that if we start centralizing for real, starting to centralize content moderation with all the best intentions we may have, we may end up with tools that could be easily misused to silence unpopular opinions or opinions that maybe somebody who has the power doesn't like. So this is really challenging because on one hand, you may want to develop tools to make the world better, to make the Internet safer and a better place for everyone, but at the same time, you may use the exact same tools for bad, I guess, purposes. Right. So I just really wanted to say it out there at the same time that I'm saying, yes, we need more kind of a standardization or unification of how we think of addressing the problem, but at the same time, we need to be very, very mindful of how the tools we may end up developing not to be kind of ending. Being misused.

Dr. Miranda Melcher (41:15)

Yeah, no, that's an incredibly important political point, philosophical point, and obviously also practical implication too. Are there any other findings or implications like that from your research that you want to make sure we discuss? Any sort of implications of what you figured out or ways forward, or ways not to go forward?

Dr. Tamar Mitts (41:35)

Yeah. So this is one of those areas that I haven't addressed fully in the book. It's kind of coming towards the end of it when I think about the implications. And what I do is I draw on the work of other fantastic scholars who have thought about this sort of, you know, growing censorship, you know, and kind of what are the dangers there? I do think that one of the implications, at least for me as a researcher in, is to better understand exactly this question, and maybe that's a new future project, is really to think about how tools for content moderation could be misused and think of that as the problem to solve of, like, how can we make sure, how can we develop systems that are not then misused? How can we safeguard that particular part of how we do things? But this is not something I currently have had enough research done to speak authoritatively on.

Dr. Miranda Melcher (42:31)

That's fair. Is that in fact your next direction or are you going somewhere else for future research?

Dr. Tamar Mitts (42:38)

It's actually related. So as a result of this book, which tends to focus mostly on the activity of non state actors, kind of what they have done online. I've became increasingly interested in what state actors are doing online. And I think this is not maybe a surprise to our listeners or anyone who kind of looks at this, but we went from an online environment 15 years ago or so, or even more when social media platforms became popular. We came from this sort of world where individual users will just post what they are doing very kind of organically during the day, sharing photos and whatever with their friends, to a very professional sort of world where the content that is being produced can be very strategically created, curated for particular sort of persuasion and influence. So I became really interested in how specifically state actors have started using the online space to influence public opinion, both, you know, domestically within their countries as well as globally. And that has led to some really new, exciting research on the role of information operations or information campaigns and kind of trying to understand who produces them for what purpose are they even successful? That's a big question we actually don't really know much about. And thinking about it, both in the context of democratic events like national elections that we had many in 2024 actually, and also in the context of conflict. So thinking about wars, the recent wars we had both between Russia and Ukraine, the war in Gaza, the war between Iran and Israel, we have a lot of conflict events where we saw many state actors engage in pretty creative information campaigns. So my new project is actually looking less at how content is kind of being moderated, although it's still definitely part of the story, but also how content is strategically produced to influence and shape political outcomes. So more on that in the future. But I'm really, really interested in this space and I would say that the mo. The more work I do on this, the more I realize, like, we need more work on this. It's like so much that we don't understand well.

Dr. Miranda Melcher (45:09)

But yeah, certainly sounds like a very interesting area to investigate. So best of luck with that.

Dr. Tamar Mitts (45:14)

Thank you.

Dr. Miranda Melcher (45:15)

In the meantime, of course, listeners can read the book we've been talking about titled Safe Havens for the Challenge of Moderating Online Extremism, published by Princeton University Press in 2025. Tamar, thank you so much for joining me on the podcast.

Dr. Tamar Mitts (45:29)

Thank you so much for having me.